Search

Mapping the spatial variability of crops is critical for precision agriculture. In this sense, remote sensing is a key technology generally dependent on the result of vegetation indices (VIs). Therefore, investigating the sensitivity of VIs and their contribution toward explaining crop variability and assisting in predicting yield is one of the pathways scientific research needs to explore. In this study, we evaluated 12 VIs with different acquisition principles in four soybean-producing fields. Using these VIs proved to be interesting to increase the performance of yield prediction models using the Randon Forest algorithm. However, simply adding VIs to the model is not enough; these VIs must aggregate information on crop variability. Some VIs are calculated based on the variation of the scene under study, and these can be an interesting option to complement the information provided by more traditional VIs, such as NDVI, assisting in predictive models, even if their direct correlation with crop yield is low in some situations. We found that using VIs groups with the same acquisition principle in isolation did not allow reaching performance of models that contained more than one principle simultaneously. In this study, the CI and TC2 indices stood out. Thus, associating VIs with different acquisition principles and, consequently, capturing different responses to variability in vegetation vigor and canopy structure is more important than the number of predictor variables itself. [ABSTRACT FROM AUTHOR]

Remote sensing offers a low-cost method for estimating yields at large spatio-temporal scales. Here, we examined the ability of Sentinel-2 satellite imagery to map field-level maize yields across smallholder farms in two regions in Oromia district, Ethiopia. We evaluated how effectively different indices, the MTCI, GCVI, and NDVI, and different models, linear regression and random forest regression, can be used to map field-level yields. We also examined if models improved by adding weather and soil data and how generalizable our models were if trained in one region and applied to another region, where no data were used for model calibration. We found that random forest regression models that used monthly MTCI composites led to the highest yield prediction accuracies (R2 up to 0.63), particularly when using only localized data for training the model. These models were not very generalizable, especially when applied to regions that had significant haze remaining in the imagery. We also found that adding soil and weather data did little to improve model fit. Our results highlight the ability of Sentinel-2 imagery to map field-level yields in smallholder systems, though accuracies are limited in regions with high cloud cover and haze. [ABSTRACT FROM AUTHOR]

Remote sensing offers a low-cost method for estimating yields at large spatio-temporal scales. Here, we examined the ability of Sentinel-2 satellite imagery to map field-level maize yields across smallholder farms in two regions in Oromia district, Ethiopia. We evaluated how effectively different indices, the MTCI, GCVI, and NDVI, and different models, linear regression and random forest regression, can be used to map field-level yields. We also examined if models improved by adding weather and soil data and how generalizable our models were if trained in one region and applied to another region, where no data were used for model calibration. We found that random forest regression models that used monthly MTCI composites led to the highest yield prediction accuracies (R2 up to 0.63), particularly when using only localized data for training the model. These models were not very generalizable, especially when applied to regions that had significant haze remaining in the imagery. We also found that adding soil and weather data did little to improve model fit. Our results highlight the ability of Sentinel-2 imagery to map field-level yields in smallholder systems, though accuracies are limited in regions with high cloud cover and haze.

Crop yield prediction and mapping are essential for crop management and decision-making. The development of unmanned aerial vehicles (UAVs) accelerates the acquisition of high-resolution imagery for the evaluation and monitoring of crop growth and development. In this study, a cost-effective yield mapping workflow was developed in a commercial setting. Two UAVs were deployed to rapidly obtain RGB and multispectral images of spinach at different growth stages in multiple spinach fields to assess crop physiological attributes in the field environment. Based on the imagery, an image processing workflow for farm field orthoimage rotation and segmentation was proposed. Then, twelve vegetation indices (VIs) were extracted from the processed images. The most robust VIs were selected by comparing the correlations among the VIs and yields and regression tree and stepwise multiple linear regression. Excess Green Index (ExG) and Normalized Difference Vegetation Index (NDVI) were identified as two of the most robust VIs for predicting spinach yield at various stages of growth. A linear regression model was developed using 45 manually harvested calibration locations and a coefficient of determination (R2) of 0.98. The root mean squared error (RMSE) was 0.19 kg/m with a mean absolute percentage error (MAPE) of 9.0 % when using a validation dataset. The produced yield maps provided the basis for a farm harvesting decision-making process and for a general crop management strategy.

Soil spatial variability is a primary contributor to within-field yield variation across farms. Spatio-temporal yield stability and variability can be assessed through multi-year yield monitor data and geostatistical techniques. Our objective was to delineate yield stability zones using multi-year yield data coupled with gross margins to plan precision conservation prescriptions. This study employed corn yield measurements from 2018, 2019, 2020 and farm economics data to compile yield stability and gross margin maps for nine Texas Blackland Prairie corn fields, and identified nonprofitable areas in each field that may be unsuitable for crop production. Yield stability zones were delineated using mean and coefficient of variation of multi-year yield maps (Zone A: high yield, stable; Zone B: high yield, unstable; Zone C: low yield, unstable; and Zone D: low yield, stable). Approximately 57% of the area in the fields was classified as unstable and, nearly 29% of the area yielded consistently below the field mean (Zone D). Gross margin for stability zones ranged from − $693 to $775/ha. Stability zones A and B generally had positive gross margins, whereas zones C and D had negative margins. Based on yield and gross margin assessment, yield stability zone D could be removed from row crop production. As a part of the Long-Term Agroecosystem Research Network Common Cropland Experiment, Zone D was removed from production (fields Y-8 and Y-13) or received reduced inputs (field SW-16 and W-13). Further study is needed to verify the farm-level economic benefits to producers and to evaluate the environmental benefits of precision conservation. [ABSTRACT FROM AUTHOR]

There is a growing body of literature that recognises the importance of UAVs in precision agriculture tasks. Currently, flowering thinning tasks in orchard management rely on the decisions derived from time-consuming manual flower cluster counting in the field by an agrotechnician. Yet it is hard to guarantee the counting accuracy due to numerous human factors. The present dataset contains UAV images during the full blooming period of an apple orchard for three consecutive years, 2018, 2019, and 2020. It is directly linked to a research article entitled “Feasibility assessment of tree-level flower intensity quantification from UAV RGB imagery: A triennial study in an apple orchard”. The data collection site was an apple orchard located at Randwijk, Overbetuwe, The Netherlands (51.938, 5.7068 in WGS84 UTM 31U). Moreover, the flower cluster number and floridity ground truth are also provided in one row from the orchard. The UAV flights were conducted with different flying altitudes, camera resolutions, and lighting conditions. This dataset aims to support researchers focussing on remote sensing, machine vision, deep learning, and image classification, and the stakeholders interested in precision horticulture and orchard management. It can be used for flowering intensity estimation and prediction, and spatial and temporal flowering variability mapping by using digital photogrammetry and 3D reconstruction.

Contemplation of potential strategies to adapt to a changing and variable climate in agricultural cropping areas depends on the availability of geo-information that is at a sufficient resolution, scale and temporal length to inform these decisions. We evaluated the efficacy of creating high-resolution, broad-scale indicators of yield from simple models that combine yield mapping data, a precision agriculture tool, with the normalised difference vegetation index (NDVI) from Landsat 5 and 7 ETM+ imagery. These models were then generalised to test its potential operationalisation across a large agricultural region (>1/2 million hectares) and the state of South Australia (>8 million hectares). Annual models were the best predictors of yield across both areas. Moderate discrimination accuracy in the regional analysis meant that models could be extrapolated with reasonable spatial precision, whereas the accuracy across the state-wide analysis was poor. Generalisation of these models to further operationalise the methodology by removing the need for crop type discrimination and the continual access to annual yield data showed some benefit. The application of this approach with past and contemporary datasets can create a long-term archive that fills an information void, providing a powerful evidence base to inform current management decisions and future on-farm land use in cropping regions elsewhere. [ABSTRACT FROM AUTHOR]

Yield maps provide a detailed account of crop production and potential revenue of a farm. This level of details enables a range of possibilities from improving input management, conducting on-farm experimentation, or generating profitability map, thus creating value for farmers. While this technology is widely available for field crops such as maize, soybean and grain, few yield sensing systems exist for horticultural crops such as berries, field vegetable or orchards. Nevertheless, a wide range of techniques and technologies have been investigated as potential means of sensing crop yield for horticultural crops. This paper reviews yield monitoring approaches that can be divided into proximal, either direct or indirect, and remote measurement principles. It reviews remote sensing as a way to estimate and forecast yield prior to harvest. For each approach, basic principles are explained as well as examples of application in horticultural crops and success rate. The different approaches provide whether a deterministic (direct measurement of weight for instance) or an empirical (capacitance measurements correlated to weight for instance) result, which may impact transferability. The discussion also covers the level of precision required for different tasks and the trend and future perspectives. This review demonstrated the need for more commercial solutions to map yield of horticultural crops. It also showed that several approaches have demonstrated high success rate and that combining technologies may be the best way to provide enough accuracy and robustness for future commercial systems. [ABSTRACT FROM AUTHOR]

Soybean crop plays an important role in world food production and food security, and agricultural production should be increased accordingly to meet the global food demand. Satellite remote sensing data is considered a promising proxy for monitoring and predicting yield. This research aimed to evaluate strategies for monitoring within-field soybean yield using Sentinel-2 visible, near-infrared and shortwave infrared (Vis/NIR/SWIR) spectral bands and partial least squares regression (PLSR) and support vector regression (SVR) methods. Soybean yield maps (over 500 ha) were recorded by a combine harvester with a yield monitor in 15 fields (3 farms) in Paraná State, southern Brazil. Sentinel-2 images (spectral bands and 8 vegetation indices) across a cropping season were correlated to soybean yield. Information pooled across the cropping season presented better results compared to single images, with best performance of Vis/NIR/SWIR spectral bands under PLSR and SVR. At the grain filling stage, field-, farm- and global-based models were evaluated and presented similar trends compared to leaf-based hyperspectral reflectance collected at the Brazilian National Soybean Research Center. SVR outperformed PLSR, with a strong correlation between observed and predicted yield. For within-field soybean yield mapping, field-based SVR models (developed individually for each field) presented the highest accuracies. The results obtained demonstrate the possibility of developing within-field yield prediction models using Sentinel-2 Vis/NIR/SWIR bands through machine learning methods. [ABSTRACT FROM AUTHOR]

Contemplation of potential strategies to adapt to a changing and variable climate in agricultural cropping areas depends on the availability of geo-information that is at a sufficient resolution, scale and temporal length to inform these decisions. We evaluated the efficacy of creating high-resolution, broad-scale indicators of yield from simple models that combine yield mapping data, a precision agriculture tool, with the normalised difference vegetation index (NDVI) from Landsat 5 and 7 ETM+ imagery. These models were then generalised to test its potential operationalisation across a large agricultural region (>1/2 million hectares) and the state of South Australia (>8 million hectares). Annual models were the best predictors of yield across both areas. Moderate discrimination accuracy in the regional analysis meant that models could be extrapolated with reasonable spatial precision, whereas the accuracy across the state-wide analysis was poor. Generalisation of these models to further operationalise the methodology by removing the need for crop type discrimination and the continual access to annual yield data showed some benefit. The application of this approach with past and contemporary datasets can create a long-term archive that fills an information void, providing a powerful evidence base to inform current management decisions and future on-farm land use in cropping regions elsewhere.

Weed maps created late in the growing season are potentially useful in regions where late maturing weeds are problematic and need to be controlled before they produce seed. The objectives of this study were to (1) spatially characterize the population dynamics of predominant weed species and apply this information into quantifying the effect of treated and untreated weed infestations on wheat (Triticum aestivum L.) yields and (2) evaluate potential herbicide savings with post-harvest site-specific treatments. Multi-year grain yield and weed data were acquired at harvest in each of four years (2015–18) within a dryland production field (9.2 ha) in eastern Oregon. Abundance of weed species (2015) and percent cover of weed species (2016, 2017 and 2018) were visually estimated on a square grid based on dividing the field into 7-m2 cells. Spatial patterns in the weed community were subject to rapid change and depended on year, crop and weed control strategy. While tumble mustard (Sisymbrium altissimum) was the most predominant and competitive species, the spatial distribution of this weed and that of other species varied each year. Tumble mustard and prickly lettuce (Lactuca serriola) were equally problematic in spring wheat and winter wheat whereas Russian thistle (Salsola tragus) was problematic in spring wheat and downy brome (Bromus tectorum) in winter wheat. Potential savings from site-specific herbicide application varied from 10 to 95% based on percentage of field infested. Weed maps at harvest are useful for studying weed dynamics, identifying potentially herbicide-resistant weeds and planning site-specific weed management. Combined with yield maps, weed maps at harvest are also useful for explaining crop yield variability that is associated with weed competition and weed control in furtherance of integrated weed management strategies. [ABSTRACT FROM AUTHOR]

Harvester-mounted yield monitor systems are increasingly used to document corn (Zea mays L.) yield. The three most commonly used spatial estimation methods to convert point data gathered with yield monitors to regular, grid-based, raster maps include nearest neighbor (NN), inverse distance weighting (IDW) and kriging. Seven spatial estimation methods (NN, IDW using 10, 20, 30 and all data points and kriging with exponential and Matérn covariance functions) were evaluated to determine the method that most accurately captures intra-field spatial variability of corn silage and corn grain yield in New York. Yield monitor data from two dairy farms and two grain operations were cleaned using Yield Editor prior to spatial analyses. The dataset included 7–10 years of data per farm for a combined 7484 ha (245 fields) of silage and 6971 ha (253 fields) of grain. Data were split into training (80%) and cross-validation datasets (remaining 20% of the data). Normalized root mean squared error (NRMSE) was used to evaluate the accuracy of the spatial estimation methods. Kriging with the Matérn covariance function resulted in the most accurate corn silage and grain yield raster maps both at the farm and field level. There were statistically significant differences in NRMSE between kriging with the Matérn isotropic covariance function and all other models for both corn silage and grain, regardless of field size, year when data were obtained or farm that supplied the data. These results are beneficial to ensure accurate and precise spatial mapping of yield products toward optimized corn growth management. [ABSTRACT FROM AUTHOR]

Normalized difference vegetation index (NDVI)-based models have been developed to derive wheat grain yields with multispectral images. In this regard, field measurements and Landsat 8 Operational Land Imager (OLI) data were used for two growing seasons to determine the relationships between NDVI and yields. The number of six statistic parameters were calculated from NDVI values to find the best agreement with actual yield data. A comparison of the results showed that sum-NDVI better matched field measurements. To compare the results of NDVI with other vegetation indices, we applied four other vegetation indices. Results indicated that estimation of wheat yields using sum-NDVI values was more accurate than estimation by sum of the four applied vegetation indices values. Also, the investigation of multi-temporal images showed that the critical time to estimate wheat yield using sum-NDVI values was the time that wheat grains were in the milky and maturity stages. [ABSTRACT FROM AUTHOR]

The phrase “soil health” describes the condition of a soil in terms of its physical, chemical, and biological properties. The relationship between soil health and crop yield is rather complex, and quantifying their spatial characteristics at field scale is often limited in the literature. To explore the significance and spatial characteristics of soil health and its indicators in crop production, field-based research was conducted in 2018 at the Blackland Research Center, Texas with the following objectives: (i) to quantify soil health indicators of the Haney Soil Health Tool across a corn field; (ii) to relate soil health status to drought-year corn yield using geospatial techniques; and (iii) to evaluate the relationship among soil health, corn yield, and soils types. We collected 218 samples from 0 to 10 cm soil depth in a 27-ha field following a 35 m × 35 m grid sampling scheme. Samples were analyzed for water extractable organic carbon (WEOC), total nitrogen, nitrate-nitrogen, ammoniacal-nitrogen, and 24-hr CO2 evolution (1-day CO2). Analyses results were used to calculate a soil health value based on Haney Soil Health Tool. The field had an average soil health value of 8.5 (±2.2), and corn yield of 2.9 t ha−1 (±1.3); however, corn yield was more variable (CV > 43%) than soil health (CV = 26.7%). Overall, soil samples with lower values of 1-day CO2, WEOC, and organic nitrogen indicated a lower soil health value, which generally corresponding to lower corn yield. Corn yield may be weakly predicted by soil health (Corn yield = −0.29 + 0.39 × Soil health value, R2 = 0.28). Soil type and rainfall exhibited measurable effects upon the calculated soil health value; corn yields were greater for Houston Black clay than Austin silt clay under drought-year conditions. The variation in drought-year corn yield can be explained by the soil health and its indicators in Texas Blackland soils. Further research is necessary to verify the relationship in differing soil types, management scenarios, and annual rainfall totals.

Crop production statistics at the field scale are scarce in African countries, limiting potential research on yield gaps as well as monitoring related to food security. This paper examines the potential of using Sentinel-2 time series data to derive spatially explicit estimates of crop production in an agroforestry parkland in central Burkina Faso. This type of landscape is characterized by agricultural fields where cereals (millet and sorghum) and legumes (cowpea) are intercropped under a relatively dense tree canopy. We measured total above ground biomass (AGB) and grain yield in 22 field plots at the end of two growing seasons (2017 and 2018) that differed in rainfall timing and amount. Linear regression models were developed using the in situ crop production estimates and temporal metrics derived from Sentinel-2 time series. We studied several important aspects of satellite-based crop production estimation, including (i) choice of vegetation indices, (ii) effectiveness of different time periods for image acquisition and temporal metrics, (iii) consistency of the method between years, and (iv) influence of intercropping and trees on accuracy of the estimates. Our results show that Sentinel-2 data were able to explain between 41 and 80% of the variation in the in situ crop production measurements, with relative root mean square error for AGB estimates ranging between 31 and 63% in 2017 and 2018, respectively, depending on temporal metric used as estimator. Neither intercropping of cereals and legumes nor tree canopy cover appeared to influence the relationship between the satellite-derived estimators and crop production. However, inter-annual rainfall variations in 2017 and 2018 resulted in different ratios of AGB to grain yield, and additionally, the most effective temporal metric for estimating crop production differed between years. Overall, this study demonstrates that Sentinel-2 data can be an important resource for upscaling field measurements of crop production in this agroforestry system in Burkina Faso. The results may be applicable in other areas with similar agricultural systems and increase the availability of crop production statistics.

The present dataset contains colour images acquired in a commercial Fuji apple orchard (Malus domestica Borkh. cv. Fuji) to reconstruct the 3D model of 11 trees by using structure-from-motion (SfM) photogrammetry. The data provided in this article is related to the research article entitled “Fruit detection and 3D location using instance segmentation neural networks and structure-from-motion photogrammetry” [1]. The Fuji-SfM dataset includes: (1) a set of 288 colour images and the corresponding annotations (apples segmentation masks) for training instance segmentation neural networks such as Mask-RCNN; (2) a set of 582 images defining a motion sequence of the scene which was used to generate the 3D model of 11 Fuji apple trees containing 1455 apples by using SfM; (3) the 3D point cloud of the scanned scene with the corresponding apple positions ground truth in global coordinates. With that, this is the first dataset for fruit detection containing images acquired in a motion sequence to build the 3D model of the scanned trees with SfM and including the corresponding 2D and 3D apple location annotations. This data allows the development, training, and test of fruit detection algorithms either based on RGB images, on coloured point clouds or on the combination of both types of data.

At an early immature growth stage of citrus, a hyperspectral camera of 369–1042 nm was employed to acquire 30 hyperspectral images in order to detect immature green fruit within citrus trees under natural illumination conditions. First, successive projections algorithm (SPA) were implemented to select 677, 804, 563, 962, and 405 nm wavebands and to construct multispectral images from the original hyperspectral images for further processing. Then, histogram threshold segmentation using NDVI of 804 and 677 nm was implemented to remove image backgrounds. Three slope parameters, calculated from the pairs 405 and 563 nm, 563 and 677 nm, and 804 and 962 nm were used to construct a classifier to identify the potential citrus fruit. Then, a marker-controlled watershed segmentation based on wavelet transform was applied to obtain potential fruit areas. Finally, a green fruit detection model was constructed according to Grey Level Co-occurrence Matrix (GLCM) texture features of the independent areas. Three supervised classifiers, logistic regression, random forest and support vector machine (SVM) were developed using texture features. The detection accuracies were 79%, 75%, and 86% for the logistic regression, random forest, and SVM models, respectively. The developed algorithm showed a great potential for identifying immature green citrus for an early yield estimation.

Yield mapping for citrus fruit is a challenging task due to factors such as varying illumination conditions, clustering, and occlusions. Mapping the yield for immature citrus fruit presents an additional challenge that is the colours of fruit and leaves are almost identical. Commonly used machine vision techniques using colour cameras become less effective for immature citrus fruit detection. This study explores a novel active thermal imaging method to tackle the problem of colour similarity between immature citrus fruit and leaves. In this study, a thermal camera was combined with a water spray system that applied water mist to citrus trees. The water mist caused temperatures of both the fruit and leaf surfaces to change but at different rates. Multiple parameters of the spray system were experimented with the goal to induce as much temperature differences as possible between fruit and leaf surfaces. The combined system was tested in a citrus grove for fruit detections. Deep learning models were built based on the active thermal imaging system and tracking and fruit counting algorithms were created to count fruit in thermal videos. A mean average precision of 87.2% was achieved by the models and an accuracy of 96% was achieved when comparing the number of fruit counted by the algorithms with the true number of fruit counted manually in the field. • An active thermal imaging system for accurate immature citrus fruit yield mapping. • Novel water spray system induces temperature differences between objects. • Various environmental and design parameters for outdoor thermal imaging assessed. [ABSTRACT FROM AUTHOR]

Crop yield maps are valuable for many applications in precision agriculture, but are often inaccessible to growers and researchers wishing to better understand yield determinants and improve site-specific management strategies. A method for mapping sub-field crop yields from remote sensing imagery could increase the availability of crop yield maps. A variation of the scalable crop yield mapping approach (SCYM, Lobell et al. in Remote Sensing of Environment 164:324–333, 2015) was developed and tested for estimating sub-field maize (Zea mays L.) yields at 10–30 m without the use of site-specific input data. The method was validated using harvester yield monitor records for 21 site-years for irrigated and rainfed fields in eastern Nebraska, USA. Prediction error ranged greatly across site-years, with relative RMSE scores of 10.8 to 38.5%, and R2 values of 0.003 to 0.37. Significant proportional bias was detected in the predictions, but could be corrected with a small amount of ground truth data. Crop yield prediction accuracies without calibration were suitable for some precision applications such as mapping relative yields and delineating management zones, but model improvements or calibration datasets are needed for applications requiring absolute yield estimates. [ABSTRACT FROM AUTHOR]

Profitability analyses of site-specific nitrogen (N) management strategies have often failed to provide satisfying reasons for adoption of precision farming technologies. However, effects of precision farming on product quality and price premiums, as well as on downside risk mitigation, are generally not taken into account. This study aimed to evaluate the comparative advantages of site-specific N management over uniform N management considering N supply on grain quality, and accordingly price premiums for wheat from a downside risk point of view. A virtual field was modelled with two subfields representing two distinctive yield zones to investigate how consideration of grain quality affects the economic potential of site-specific N management under temporally varying N mineralization and changing price patterns of wheat. Moreover, the extent was investigated to which site-specific N management can have a risk-reducing effect on economic shortfalls compared to uniform N management. Two site-specific N management options were assessed: variable N rate application using yield mapping and N sensor for real-time proximal sensing. Results indicated that even though crop yields were only slightly higher, higher expected protein contents of grains could be achieved with site-specific N management options compared to uniform N management. Baking wheat quality was secured to a greater extent with site-specific N management options. Higher average grain quality improved the economic benefits due to price premiums. A risk-reducing effect was observed with the site-specific N management by maintaining the baking wheat quality with a higher probability. Higher economic returns mostly compensated the additional costs for the precision farming technologies in the lower tail of the probability distribution and, thus, site-specific N management did not show any substantial disadvantage on downside risk as compared to uniform N management. [ABSTRACT FROM AUTHOR]

Accurate measurements of maize yields at field or subfield scales are useful for guiding agronomic practices and investments and policies for improving food security. Data on smallholder maize systems are currently sparse, but satellite remote sensing offers promise for accelerating learning about these systems. Here we document the use of Google Earth Engine (GEE) to build "wall-to-wall" 10 m resolution maps of (i) cropland presence, (ii) maize presence, and (iii) maize yields for the main 2017 maize season in Kenya and Tanzania. Mapping these outcomes at this scale is extremely challenging because of very heterogeneous landscapes, lack of cloud-free satellite imagery, and the low quantity of quality ground-based data in these regions. First, we computed seasonal median composites of Sentinel-1 radar backscatter and Sentinel-2 optical reflectance measures for each pixel in the region, and used them to build both crop/non-crop and maize/non-maize Random Forest (RF) classifiers. Several thousand crop/non-crop labels were collected through an in-house GEE labeler, and thousands of crop type labels from the 2015–2017 growing seasons were obtained from various sources. Results show that the crop/non-crop classifier successfully identified cropland with over 85% out-of-sample accuracy in both countries, with Sentinel-1 being particularly useful for prediction. Among the cropped pixels, the maize/non-maize classier had an accuracy of 79% in Tanzania and 63% in Kenya. To map maize yields, we build on past work using a scalable crop yield mapper (SCYM) that utilizes simulations from a crop model to train a regression that predicts yields from observations. Here we advance past approaches by (i) grouping simulations by Global Agro-Environmental Stratification (GAES) zones across the two countries, in order to account for landscape heterogeneity, (ii) utilizing gridded datasets on soil and sowing and harvest dates to setup model simulations in a scalable way; and (iii) utilizing all available satellite observations during the growing season in a parsimonious way by using harmonic regression fits implemented in GEE. SCYM estimates were able to capture about 50% of the variation in the yields at the district level in Western Kenya as measured by objective ground-based crop cuts. Finally, we illustrated the utility of our yield maps with two case studies. First, we document the magnitude and interannual variability of spatial heterogeneity of yields in each district, and how it varies for different parts of the region. Second, we combine our estimates with recently released soil databases in the region to investigate the most important soil constraints in the region. Soil factors explain a high fraction (72%) of variation in predicted yields, with the predominant factor being soil nitrogen levels. Overall, this study illustrates the power of combining Sentinel-1 and Sentinel-2 imagery, the GEE platform, and advanced classification and yield mapping algorithms to advance understanding of smallholder agricultural systems. • National maps of crop and maize areas are generated at 10 m resolution in Africa. • Maize yields are mapped accurately without any ground calibration. • Independent soil data can explain over 70% of variability in yield estimates. • Soil nitrogen and organic carbon exhibit the strongest association with yields. • Google Earth Engine and Sentinel data offer new opportunities for mapping agriculture. [ABSTRACT FROM AUTHOR]

Agricultural activities conducted in the Great Rift Valley of Kenya show a significant decline of productivity levels. This phenomenon is mainly related to limited availability of water resources, lack of supporting irrigation, and harvesting techniques ineffectiveness. Production risks reduction is closely related with a better use of water resources and a better understanding of the effects resulting from the multiple interactions between climate, agricultural vegetation, soil type, and crops management techniques. In this paper, a remote and automatic agricultural monitoring system is presented as an effective alternative to the most traditional in situ measurements and observations. We investigated the use of phenological information extracted from satellite imagery combined with crop calendar and supported by agro-ecological zoning (AEZ) in accurate crop classification and monitoring. Vegetation indices extracted from Landsat 8 imagery are capable to track the vegetation development through the year, then phenological profiles can be extracted and implemented into a multitemporal automatic classification process to detect agricultural areas and to discriminate among different crop species. The phenological profiles extracted by satellite imagery are compared with crop calendar data compiled by FAO for the area of interest. The classification procedure is supported by AEZs based on crop modeling and environmental matching procedures in order to identify crop-specific environmental limitations under assumed levels of inputs and management conditions. The FAO crop water productivity model AquaCrop is calibrated for wheat and maize yield mapping in the central highland of Kenya, handling both environmental and phenological data. The combined use of phenological data and AEZs results in a robust methodology with a classification overall accuracy of 91.35%. A good model performance is obtained relative to yield predictions, with R of 0.69 and 0.72. [ABSTRACT FROM AUTHOR]

A survey of Australian grain growers was conducted to gauge grower attitudes to crop and soil sensing and their role in nitrogen fertilizer management. The technologies considered were yield monitors, remote and proximal crop sensing, high resolution soil survey, soil moisture sensing and digital elevation models (DEM). Whereas Australian grain growers have readily adopted machine guidance and autosteer, and a majority have access to yield monitoring, the rate of use of many crop and soil sensors remains comparatively low. However, the survey results suggest a positive effect on sensor adoption through present use of yield mapping. Access to yield maps was significantly associated with the use of remotely sensed imagery, high resolution soil survey, soil moisture sensing, DEM and variable rate application of fertilizers and/or soil amendments. There is some support for proximal crop sensing, albeit with low present rates of use; the use of soil water sensors and DEM is presently very low. For the further development of precision agriculture (PA), the results make clear that expending effort in enhancing the adoption and use of yield maps would be valuable as a lever to gain 'buy-in' from growers to sensing and PA more broadly. Since growers use many sources of information to support fertilizer decision-making, any new fertilizer decision aid needs to establish a point of difference from, but be complementary to, existing tools. One way of achieving this would be to use sensors, supported by locally derived algorithms, as a key input to fertilizer decision support. [ABSTRACT FROM AUTHOR]

Cloud computing and freely available, high-resolution satellite data have enabled recent progress in crop yield mapping at fine scales. However, extensive validation data at a matching resolution remain uncommon or infeasible due to data availability. This has limited the ability to evaluate different yield estimation models and improve understanding of key features useful for yield estimation in both data-rich and data-poor contexts. Here, we assess machine learning models’ capacity for soybean yield prediction using a unique ground-truth dataset of high-resolution (5 m) yield maps generated from combine harvester yield monitor data for over a million field-year observations across the Midwestern United States from 2008 to 2018. First, we compare random forest (RF) implementations, testing a range of feature engineering approaches using Sentinel-2 and Landsat spectral data for 20- and 30-m scale yield prediction. We find that Sentinel-2-based models can explain up to 45% of out-of-sample yield variability from 2017 to 2018 (r2 = 0.45), while Landsat models explain up to 43% across the longer 2008–2018 period. Using discrete Fourier transforms, or harmonic regressions, to capture soybean phenology improved the Landsat-based model considerably. Second, we compare RF models trained using this ground-truth data to models trained on available county-level statistics. We find that county-level models rely more heavily on just a few predictors, namely August weather covariates (vapor pressure deficit, rainfall, temperature) and July and August near-infrared observations. As a result, county-scale models perform relatively poorly on field-scale validation (r2 = 0.32), especially for high-yielding fields, but perform similarly to field-scale models when evaluated at the county scale (r2 = 0.82). Finally, we test whether our findings on variable importance can inform a simple, generalizable framework for regions or time periods beyond ground data availability. To do so, we test improvements to a Scalable Crop Yield Mapper (SCYM) approach that uses crop simulations to train statistical models for yield estimation. Based on findings from our RF models, we employ harmonic regressions to estimate peak vegetation index (VI) and a VI observation 30 days later, with August rainfall as the sole weather covariate in our new SCYM model. Modifications improved SCYM’s explained variance (r2 = 0.27 at the 30 m scale) and provide a new, parsimonious model.

Precision agriculture is a management approach for sustainable agriculture. It can be applied even in small fields. It aims to optimize inputs, improve profits, and reduce adverse environmental impacts. In this study, a series of measurements were conducted over three growing seasons to assess variability in a 0.55 ha pear orchard located in central Greece. Soil ECa was measured using EM38 sensor, while soil samples were taken from a grid 17 × 8 m and analysed for texture, pH, P, K, Mg, CaCO3, and organic matter content. Data analysis indicated that most of the nutrients were at sufficient levels. Soil and yield maps showed considerable variability while fruit quality presented small variations across the orchard. Yield fluctuations were observed, possibly due to climatic conditions. Prescription maps were developed for nitrogen variable rate application (VRA) for two years based on the replacement of the nutrients removed by the crop. VRA application resulted in 56% and 50% reduction of N fertiliser compared to uniform application.

Early detection and counting of immature green citrus fruit using computer vision can help growers produce a predictive yield map which could be used to adjust management practices during the fruit maturing stages. However, such detecting and counting is difficult because of varying illumination, random occlusion and color similarity with leaves. An immature fruit detection algorithm was developed with the aim of identifying and counting fruit in a citrus grove under varying illumination environments and random occlusions using images acquired by a regular red-green-blue (RGB) color camera. Acquired citrus images included front-lighting and back-lighting illumination conditions. The Retinex image enhancement algorithm and the two-dimensional discrete wavelet transform were used for image illumination normalization. Color-based K-means clustering and circular hough transform (CHT) were applied in order to detect potential fruit regions. A Local Binary Patterns feature-based Adaptive Boosting (AdaBoost) classifier was built for removing false positives. A sub-window was used to scan the difference image between the illumination-normalized image and the resulting image from CHT detection in order to detect small areas and partially occluded fruit. An overall accuracy of 85.6% was achieved for the validation set which showed promising potential for the proposed method. [ABSTRACT FROM AUTHOR]

Highlights • An instrumented picking cart is developed for strawberry yield monitoring. • The yield-monitoring device is equipped with load cells, an RTK GPS, and an IMU. • A comprehensive calibration process is performed for the picking cart's load cells. • A yield map for a 300 m 2 plot of a strawberry field is generated. • The mean prediction accuracy of the load cells is 4.8 %. Abstract Strawberries in California have a $2 billion direct economic impact on the state; however, they are currently produced based only on uniform field management techniques. Creating yield maps of strawberries could allow for variable-rate and site-specific applications of inputs, which could improve productivity and reduce environmental pollution. This paper presents the development and application of an instrumented strawberry-picking cart for yield mapping. During manual harvest of strawberries planted on raised beds, pickers walk inside the furrows, pick fruit from the beds on both sides, and deposit them into a tray placed on a picking cart. A 'smart' picking cart, similar to the standard carts, has been designed and instrumented with several types of sensors including load cells, a real-time kinematic global positioning system (RTK GPS), a microcontroller, and an inertial measurement unit (IMU). This instrumented cart serves two purposes: to work in sync with tray-transporting robots during robot-aided strawberry harvesting, and to create yield maps of strawberry fields. A yield map for an approximately 300 m 2 plot of a strawberry field in Salinas, California, was generated after the cart was calibrated. During the yield-monitoring experiment, 13.5 trays, each with a capacity of about 4.2 k g of strawberries, were filled with fruit. The mean prediction accuracy of the mass of full trays measured by the load cells was calculated to be 4.8 %. [ABSTRACT FROM AUTHOR]

Citrus fruit detection is one of the most important and challenging steps in citrus yield mapping. The distinct color differences between the ripe fruit and leaves allowed previously-described imaging-based methods to achieve good results. However, immature green citrus fruit detection, which aims to provide valuable information for citrus yield mapping at earlier stages is much more difficult because the fruit and leaf colors are very similar. This study combines color and thermal images for immature green fruit detections. Experiments identified optimal conditions for thermal imaging. A multimodal imaging platform was built to integrate color and thermal cameras. A novel image registration method was developed for combining color and thermal images and matching fruit in both images which achieved pixel-level accuracy. A new Color-Thermal Combined Probability (CTCP) algorithm was created to effectively fuse information from the color and thermal images to classify potential image regions into fruit and non-fruit classes. Algorithms were also developed to integrate image registration, information fusion and fruit classification and detection into a single step for real-time processing. An increase in recall rate from 78.1% when using only color images to 90.4% after fusing the color and thermal images was obtained at similar precision rates, and an increase in precision rate from 86.6% to 95.5% was obtained at similar recall rates. The fusion of the color and thermal images effectively improved immature green citrus fruit detection. [ABSTRACT FROM AUTHOR]

Assessing the spatial variation of soil and crop properties is the basis for site specific management of crop practices in precision agriculture applications. To this aim, proximal and remote spectral vegetation indices are increasingly replacing soil analysis. In this study the spatial variation of soil properties, proximal and remote spectral vegetation indices were compared in a winter wheat (Triticum aestivum L.) crop grown in a 4.15 ha field in northern Italy. Soil analysis (particle size distribution, pH, carbonates, C, total N, available P, exchangeable cations and electrical conductivity) was geo-referentially carried out; the proximal indices chlorophyll content by N-Tester and normalised difference vegetation index through GreenSeeker were determined in three dates during stem elongation; the remote indices PurePixelTM chlorophyll index and PurePixelTM vegetation index were determined through the Landsat 8 satellite in three dates during the same wheat stage. Dry biomass yield (DBY), grain yield (GY) and yield components were determined at harvest. Soil, proximal and remote data were submitted to principal component analysis (PCA), and the retained PCs were clustered to delineate areas at low, intermediate and high yield potential, based on soil parameters (CLUsp), proximal (CLUpi), and remote vegetation indices (CLUri). DBY and GY were significantly correlated with several soil parameters and vegetation indices. Spatial distribution of soil and crop data consistently depicted a low performing area (GY<3 Mg ha-1) and a high performing one (GY>8 Mg ha-1). CLUsp determined a lower GY difference between low and high performing area (+60%), compared to CLUpi and CLUri (almost +100%). In CLUsp and CLUpi the low and high performing area were of similar size (25 and 29% for the two respective areas in CLUsp; 25 and 33% in CLUpi), whereas in CLUri they were quite different (16 and 46%). Lastly, yield potential levels determined by vegetation indices (CLUpi and CLUri) exhibited a better degree of agreement with DBY and GY levels, than soil parameters (CLUsp). In exchange for this, the above referred soil parameters are quite consistent in time, allowing soil data to be used for more years. On concluding, PCA followed by clustering resulted in a robust delineation of field areas at different yield potential. This is the premise for developing research driven strategies of practical use. [ABSTRACT FROM AUTHOR]

Yield mapping is a subject of research in (precision) agriculture and one of the primary concerns for farmers as it forms the basis of their income and has implications for subsidies and taxes. The presented approach involves deployment of field harvesters equipped with sensors that provide more detailed and spatially localized values than merely a sum of yields for the whole plot. The measurements from such sensors need to be filtered and subject to further processing, including interpolation, to facilitate follow-up interpretation. This paper aims to identify the relative differences between interpolations from (1) (field) measured data, (2) measured data that were globally filtered, and (3) measured data that were globally and locally filtered. All the measured data were obtained at a fully operational farm and are considered to represent a natural experiment. The revealed spatial patterns and recommendations regarding global and local filtering methods are presented at the end of the paper. Time investments into filtering techniques are also taken into account.

The availability of big data in agriculture, enhanced by free remote sensing data and on-board sensor-based data, provides an opportunity to understand within-field and year-to-year variability and promote precision farming practices for site-specific management. This paper explores the performance in durum wheat yield estimation using different technologies and data processing methods. A state-of-the-art data cleaning technique has been applied to data from a yield monitoring system, giving a good agreement between yield monitoring data and hand sampled data. The potential use of Sentinel-2 and Landsat-8 images in precision agriculture for within-field production variability is then assessed, and the optimal time for remote sensing to relate to durum wheat yield is also explored. Comparison of the Normalized Difference Vegetation Index(NDVI) with yield monitoring data reveals significant and highly positive linear relationships (r ranging from 0.54 to 0.74) explaining most within-field variability for all the images acquired between March and April. Remote sensing data analyzed with these methods could be used to assess durum wheat yield and above all to depict spatial variability in order to adopt site-specific management and improve productivity, save time and provide a potential alternative to traditional farming practices.

This study aimed to compare standard and precision nitrogen (N) fertilization with variable rate technology (VRT) in winter wheat (Triticum aestivum L.) by combining data of NDVI (Normalized Difference Vegetation Index) from the Sentinel 2 satellite, grain yield mapping, and protein content. Precision N rates were calculated using simple linear models that can be easily used by non-specialists of precision agriculture, starting from widely available Sentinel 2 NDVI data. To remove the effects of not measured or unknown factors, the study area of about 14 hectares, located in Central Italy, was divided into 168 experimental units laid down in a randomized design. The first fertilization rate was the same for all experimental units (30 kg N ha−1). The second one was varied according to three different treatments: 1) a standard rate of 120 kg N ha−1 calculated by a common N balance; 2) a variable rate (60−120 kg N ha−1) calculated from NDVI using a linear model where the maximum rate was equal to the standard rate (Var-N-low); 3) a variable rate (90−150 kg N ha−1) calculated from NDVI using a linear model where the mean rate was equal to the standard rate (Var-N-high). Results indicate that differences between treatments in crop vegetation index, grain yield, and protein content were negligible and generally not significant. This evidence suggests that a low-N management approach, based on simple linear NDVI models and VRT, may considerably reduce the economic and environmental impact of N fertilization in winter wheat.

Developing regionally applicable models for timely mapping of rice yield potential still faces challenges like the poor accuracy of yield prediction at early stages, and the scale mismatch between ground samples and satellite image pixels. Here, we explored the potential of using high-resolution multispectral images and texture features derived from unmanned aerial system (UAS) to fill these gaps. Rice field experiments were carried out in eastern China during 2017, 2018, and 2020 to acquire UAS (eBee-SQ) and satellite (Sentinel-2) images at jointing stage as well as the rice final yield. High spatial-resolution (HSR, 0.1 m) eBee-SQ images were used to extract both textures and vegetation indices (VIs), the resampled medium spatial-resolution (MSR, 20 m) eBee-SQ and Sentinel-2 images were only used to extract VIs. Firstly, using experimental data from 2017 and 2018, five kinds of yield prediction models at jointing stage were developed with: [ⅰ] VIs derived from MSR Sentinel-2 images and linear regression (LR) (least-RMSE = 1.68 t ha−1), [ⅰi] VIs derived from MSR eBee-SQ images and LR (least-RMSE = 1.482 t ha−1), [ⅱi] VIs derived from HSR eBee-SQ images and LR (least-RMSE = 1.371 t ha−1), [iv] VI-texture mixed indices derived from HSR eBee-SQ images and LR (least-RMSE = 1.082 t ha−1), and [v] random forest (RF) regression combining VIs and textures derived from HSR eBee-SQ images (least-RMSE = 0.997 t ha−1). These results showed that yield prediction models developed by both MSR and HSR eBee-SQ data performed better than by MSR Sentinel-2 data. Secondly, using optimal models derived from HSR eBee-SQ data and in-season HSR eBee-SQ data, different farm-scale HSR yield potential maps were generated. These yield maps were processed to Sentinel-2 resolution (20 m) to relate with Sentinel-2 VI for building integrated yield prediction models (IPMs) based on linear and quadratic regression. IPMs were then used to map town-scale rice yield potential by importing Sentinel-2 VI from jointing stage, and were compared with simple yield prediction model (SPM) that directly used Sentinel-2 VI. The independent validation results using yield data collected in 2020 suggested that all of the IPMs were more accurate (RMSE = 0.64 – 0.74 t ha−1) than the SPM (RMSE = 0.96 t ha−1). Besides, IPMs involving eBee-SQ derived textures and VIs showed higher accuracy in mapping yield potential than IPM involving eBee-SQ derived VI only. Overall, this study provides a novel approach to accurately mapping the regional rice yield potential through the integration of satellite and UAS data. • Combining VI and texture improved yield prediction accuracy at rice jointing stage. • Textures from red-edge band were more conducive to yield prediction in RF model. • UAS-predicted yield map was upscaled and correlated with satellite VI to build IPM. • IPM improved accuracy for mapping town-level yield potential compared with SPM. • UAS served as an intermediate source to fill gaps between ground and satellite data. [ABSTRACT FROM AUTHOR]

There is a growing body of literature that recognises the importance of UAVs in precision agriculture tasks. Currently, flowering thinning tasks in orchard management rely on the decisions derived from time-consuming manual flower cluster counting in the field by an agrotechnician. Yet it is hard to guarantee the counting accuracy due to numerous human factors. The present dataset contains UAV images during the full blooming period of an apple orchard for three consecutive years, 2018, 2019, and 2020. It is directly linked to a research article entitled "Feasibility assessment of tree-level flower intensity quantification from UAV RGB imagery: A triennial study in an apple orchard". The data collection site was an apple orchard located at Randwijk, Overbetuwe, The Netherlands (51.938, 5.7068 in WGS84 UTM 31U). Moreover, the flower cluster number and floridity ground truth are also provided in one row from the orchard. The UAV flights were conducted with different flying altitudes, camera resolutions, and lighting conditions. This dataset aims to support researchers focussing on remote sensing, machine vision, deep learning, and image classification, and the stakeholders interested in precision horticulture and orchard management. It can be used for flowering intensity estimation and prediction, and spatial and temporal flowering variability mapping by using digital photogrammetry and 3D reconstruction., Competing Interests: The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper., (© 2023 The Authors. Published by Elsevier Inc.)